Augmenting Around-Device Interaction by Geomagnetic Field Built-in Sensor Utilization.

Area of Science:

• Human-Computer Interaction
• Mobile Computing
• Sensor Technology

Background:

• Traditional mobile device interaction relies heavily on touch-based gestures.
• Exploring alternative input methods is crucial for enhancing user experience and accessibility.
• Mobile devices possess underutilized built-in sensors, such as magnetic field sensors.

Purpose of the Study:

• To investigate magnet-based Around-Device Interaction techniques for mobile devices.
• To develop novel input methods that augment interaction without traditional touch gestures.
• To explore the use of magnetic field sensors for gesture-free mobile interaction.

Main Methods:

• Utilizing a smartphone's built-in magnetic field sensor.
• Applying magnetic field fingerprinting, curve-fitting, and machine learning algorithms.
• Developing proof-of-concept applications for text entry and free-move pointing.

Main Results:

• Achieved a text entry speed of nine words per minute using discrete magnet positioning.
• Demonstrated that neural networks outperform curve fitting for free-move pointing.
• Obtained a mean absolute error of 3 mm for pointer position using neural networks.

Conclusions:

• Magnet-based Around-Device Interaction effectively expands mobile device interaction language and space.
• Machine learning, particularly neural networks, offers a viable and accurate approach for magnet-based mobile input.
• This research presents a hardware-free method for enhancing mobile device interactivity.